Web Note A
Assembly of microarrays for genome-wide measurement of DNA copy number by CGH
Antoine M. Snijders
1,2
, Norma Nowak
4
, Richard Segraves
1
, Stephanie Blackwood
1,2
, Nils
Brown
1
, Jeffrey Conroy
4
, Greg Hamilton
1
, Anna Katherine Hindle
1, 2
, Bing Huey
1
, Karen
Kimura
1
, Sindy Law
1,2
, Ken Myambo
1
, Joel Palmer
1, 2
, Bauke Ylstra
1, 2
, Jingzhu Pearl Yue
1
, Joe
W. Gray
1, 3
, Ajay N. Jain
1,2,3
, Daniel Pinkel
1,3
and Donna G. Albertson
1,2,3
1
Comprehensive Cancer Center,
2
Cancer Research Institute and
3
Department of Laboratory
Medicine, University of California San Francisco, San Francisco, CA 94143-0808 and
4
Roswell
Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263
Address for correspondence:
Donna G. Albertson
Cancer Research Institute
Box 0808
University of California San Francisco
San Francisco, CA 94143-0808
Tel. (415) 502-8463
Fax. (415) 476-8218
Email: albertson@cc.ucsf.edu

Methods
Specimens . We obtained GM03563, GM00143, GM05296, GM07408, GM01750, GM03134,
GM13330, GM03576, GM01535, GM07081, GM02948, GM04435, GM10315, GM13031 and
GM01524 cell strains from the NIGMS Human Genetics Cell Repository (Coriell Institute for
Medical Research), and cultured them under the recommended conditions. We cultured cell lines
BT474, HCT116, T47D, MPE600, COLO320, SW837, MDA-MB-231, MDA-MB-453 and HT29 under the recommended conditions. To isolate DNA from confluent cell cultures, we collected cells from T75 flasks and then incubated them overnight at 55

C in a 3 ml solution containing 0.01 M
Tris, pH 7.5, 0.001 M EDTA, pH 8, 0.5% SDS and 0.1
 g/
 l proteinase K. We added 1 ml of saturated NaCl solution and 10 ml of ethanol to the cell lysate and collected the DNA by spooling.
We removed excess liquid and then dissolved the DNA in 1 ml of H
2
O. We isolated DNA from breast tumor specimens as described previously
1
.
Genomic clones . We selected the majority of the clones for the arrays from the set of cytogenetically mapped BACs reported previously
4
and obtained these clones from the Roswell Park
Cancer Institute. We obtained additional clones mapping near the telomeres of each chromosome
5 and clones containing certain named genes from J. Flint, D. Ledbetter, C. Lese and Vysis, Inc. We included P1 clones used previously on arrays of chromosome 20 (ref. 1,2). We also used additional clones from the DuPont A library, including RMC01P052 (163D7), RMC01P057 (213C4),
RMC07P014 (548H7), RMC07P025 (1128F4), RMC07P028 (252B4) and RMC07P038
(1429E1.c2), as well as RMC17P041 (75D7), RMC17P069 (88H6) and RMCXP002 (124A3) from the DuPont B library. The array provides only minimal coverage of the Y chromosome, since only one clone unique to the Y chromosome is included on the array in addition to the telomeric clones that are shared between the X and Y chromosomes.
1

Isolation of BAC/P1 DNA . We inoculated 0.1
 l of a glycerol stock of BAC or P1 containing bacteria into 5 ml LB media containing chloramphenicol (10
 g/ml) or kanamycin (50
 g/ml) for
BACs or P1s, respectively and incubated the cultures overnight (~16 h) at 37

C with agitation at
225 rpm. We prepared larger cultures for DNA isolation, by inoculating 25 ml
LB media, containing the appropriate antibiotic with 200
 l of the previously grown overnight culture and incubating in a shaking incubator at 37

C at 225 rpm overnight (~16 h). We monitored bacterial growth by measuring the OD of a 1:10 dilution, which preferably ranged between 0.25 and 0.35 at a wavelength of 600 nm. We used the Qiagen Plasmid Mini kit to isolate DNA, following a modified version of the Qiagen Plasmid Mini Purification protocol.
We transferred the 25 ml cultures into 50 ml tubes and centrifuged them for 15 min at 4000 rcf at
4

C to pellet the bacteria. We re-suspended the pellet in Qiagen buffer P1 (1.5 ml), containing
RNase A at the manufacturer’s recommended concentration and then added buffer P2 (1.5 ml).
We mixed the tubes extremely gently by inversion and incubated at room temperature for 5 min.
Then, we added buffer P3 (1.5 ml), again mixed very gently by inversion and incubated on ice for 10 min. We inverted each tube once and centrifuged at 4000 rpm at 4

C until the supernatant became clear (45 to 60 min). We filtered the supernatant through a 35-micron nylon mesh prior to loading onto Qiagen-tip 20 columns, which had been equilibrated following the manufacturer’s protocol. We followed the manufacturer’s protocol for the DNA isolation steps with the exception that the elution buffer QF was heated to 65

C before it was added to the column. After elution, we added 0.56 ml isopropanol to the DNA and incubated overnight at 4

C. We collected the DNA pellet by centrifugation for 45 min at 14000 rpm at 4

C. After aspirating the supernatant, we allowed the pellet to dry in air for at most 30 min before resuspending the DNA in 50
 l of H
2
O. We determined the DNA concentration using a
2

fluorometer. The yield typically ranged between 60-100 ng/
 l (i.e. 3-5
 g per isolation). To ascertain DNA purity, we digested each BAC (200 ng) with Hin dIII and electrophoresed the digest through a 0.75% agarose gel. We discarded DNA preparations that showed a significant contamination with host bacterial DNA, seen as a background smear of degraded DNA in the gel.
Preparation of BAC/P1 DNA representations by ligation-mediated PCR . We digested the DNA with Mse I by incubating overnight at 37

C in a 5
 l reaction containing 1.5
 l DNA (20 to 600 ng),
0.4 U/
 l Mse I (New England Biolabs), and 0.4x
One-Phor-All-Buffer-Plus (Amersham). For ligation of adapters, we diluted 1
 l of each digest to a final concentration of 1 ng/
 l with H
2
O and then mixed 1 ng of the digested DNA with 0.5
 l One-Phor-All-Buffer-Plus (10x, Amersham), 0.5
 l of 100

M Primer 1 (TAACTAGCATGC), 0.5
 l of 100
 M Primer 2 (5’ aminolinker
AGTGGGATTCCGCATGCTAGT) and 5.5
 l H
2
O. We incubated the mixture at 65

C for 1 min after which we ramped the temperature down to 15

C with a ramp-speed of 1.3

C per minute.
When the temperature reached 15

C, we added 1
 l of 10 mM ATP and 1
 l T4-DNA ligase (5
U/
 l, Gibco BRL) and continued incubation at 15

C overnight. To initiate the first round of PCR amplification, we added a 40
 l mixture, consisting of 3
 l PCR Buffer 1 (Expand Long Template
PCR System, Roche), 2
 l of a mixture of each nucleotide (10 mM) and 35
 l H
2
O to the ligation reaction. Before we started the PCR program, we incubated the reaction at 68

C for 4 min and then added 1
 l DNA polymerase (3.5 U/
 l, Expand Long Template PCR System, Roche). We carried out thermal cycling in a Perkin-Elmer Gene Amp PCR System 9700 block as follows: 94

C for 40 s, 57

C for 30 s, 68

C for 1.25 min for 14 cycles, followed by, 94

C for 40 s, 57

C for 30 s, 68

C for 1.75 min for 34 cycles and 94

C for 40 s, 57

C for 30 s and 68

C for 5 min for the final cycle.
3

We electrophoresed 3.5
 l of the PCR product through a 1% agarose gel to determine the size range of the amplified DNA, which ideally ranged from 100 to 2000 bp. To make the DNA for spotting on the arrays, we carried out a second round of amplification in a 100
 l reaction containing 1
 l of the primary PCR product, 4

M Primer 2, TAQ-buffer II (1x; Perkin Elmer), 0.2 mM dNTP mix, 5.5 mM MgCl
2
(Perkin Elmer), 2.5 U Amplitaq Gold (Perkin Elmer) and H
2
O. We carried out an initial incubation at 95

C for 10 min in a MJ Research Peltier Thermal Cycler 225, followed by 95

C for
30 s, 50

C for 30 s and 72

C for 2 min for 45 cycles and finally 7 min at 72

C. This reaction yields ~10
 g of DNA, with each fragment containing a 5’ amino linker.
Preparation of DNA spotting solutions . We evaporated the amplification reaction (100
 l) to a final volume of 50
 l by incubation at 45

C in a hybridization oven (Techne, Hybridizer HB-1D) for approximately 75 min and then added 2.5 volumes of ice-cold ethanol and 0.1 volumes of 3 M
NaOAc to precipitate the DNA. (The use of ethanol precipitation proved superior to isopropanol precipitation.) We inverted the tubes and chilled them at –20

C for 15 min before collecting the precipitate by centrifugation at 1699 rcf for 90 min. We washed the pellets with 70% ethanol (150
 l) and collected them by centrifugation at 1699 rcf for 45 min. We air-dried the pellets for approximately 60 to 90 min and then re-suspended them in 12
 l of 20% DMSO in H
2
O (~0.8
 g/
 l). We transferred the DNA solutions into 864 well microtitre plates for robotic arraying.
Previously, we prepared DNA for spotting in 80% DMSO and 0.3
 g/
 l nitrocellulose. However, we subsequently evaluated spotting solutions containing various concentrations of dimethyl formamide, formamide or DMSO, with or without nitrocellulose. We determined that nitrocellulose was not required for spotting DNA prepared by PCR using primers with 5’amino linkers and that spotting solutions made with 20% DMSO performed well. We have also generated representations of large insert clones for arraying by using degenerate oligonucleotide primed PCR and by
4

amplification of mixtures of subclones from BACs, but found the ligation-mediated PCR procedure to be superior.
Array printing . We used a custom built printer, employing a 4 x 4 array of quartz capillary tubes spaced on 3 mm centers to print ~70-100 μm diameter spots on 130 μm centers. We printed each
DNA solution in triplicate to create an array of ~7500 elements in a 12 mm square area. We printed the arrays on chromium coated microscope slides (PTI or Nanofilm) for these studies, but also routinely print the arrays on glass slides (Corning GAPs). We allowed the printed slides to dry overnight, then exposed the slides to UV light (65 mJ) in a UV Stratalinker 2400 (Stratagene). We hybridized to these slides without additional treatment, except for the pre-hybridization slide blocking described below. We found no indication of DNA loss from the spots at any stage of the procedure when we performed hybridization in formamide buffers at 37 °C.
DNA labeling . We labeled DNA by nick translation or random priming. For nick translation, we used a 400
 l reaction containing 10
 g DNA, 50 mM Tris, pH 7.6, 5 mM MgCl
2
, 0.02 mM each dATP, dTTP and dGTP, 0.2 mM Cy3- or Cy5-dCTP (Amersham), 0.2 U DNA Polymerase I (Gibco
BRL), and 30
 l DNAseI/PolI enzyme mix (Gibco BRL). We incubated the reaction at 15

C for 60 min after which we stopped the reaction by incubation at 70

C for 10 min. We removed unincorporated nucleotides using a Sephadex G-50 spin column. We labeled genomic DNA by random priming in a 100
 l reaction containing 0.003 - 0.6
 g DNA, 1x random primers solution
(BioPrime DNA Labeling System, Gibco BRL), 1 mM Tris, pH 7.6, 0.1 mM EDTA, 0.2 mM each of dATP, dTTP and dGTP, 0.1 mM dCTP, 0.4 mM Cy3 or Cy5-dCTP (Amersham) and 160 U Klenow fragment (BioPrime DNA Labeling System, Gibco BRL). We incubated the DNA with the random primers solution at 100

C for 10 min in a total volume of 84
 l, prior to adding the other reagents
5

and then incubated the100
 l reaction overnight at 37

C. We removed unincorporated nucleotides using a Sephadex G-50 column.
6

Hybridization . We combined test and reference DNAs (~2
 g of input genomic DNA for nick translation or ~0.6
 g of input genomic DNA for random prime labeling) with Cot-1 DNA (80-
100
 g; Gibco BRL) and precipitated them with ethanol. We collected the precipitate by centrifugation and allowed the precipitate to dry in air for 10 min before re-dissolving it in a 50
 l hybridization mixture containing 50% formamide, 2 x SSC, 10% dextran sulfate, 4% SDS and
500
 g yeast tRNA, pH 7. We incubated the hybridization mixture at 70

C for 10-15 min to denature the DNA and subsequently continued incubation at 37

C for 60 min to allow blocking of repetitive sequences. We applied a ring of rubber cement closely around the array to form a well, into which we added 50
 l of slide blocking solution containing 500
 g salmon sperm
DNA in 50% formamide, 2 x SSC, 10% dextran sulfate and 4% SDS, pH 7. We created an airtight hybridization chamber by placing a silicone gasket (PGC Scientific) around the array and rubber cement ring, placing a slide on top of it and clamping with binder clips. After a 30 min incubation at room temperature, we opened the chamber, removed approximately three-quarters of the blocking solution, added the denatured and re-annealed hybridization mixture and then resealed the chamber. We placed the arrays on a slowly rocking table (~1 rpm) at 37

C to allow hybridization to occur over 16-72 h. After hybridization, we rinsed off the excess hybridization fluid with PN buffer (PN: 0.1 M sodium phosphate, 0.1% nonidet P40, pH 8), then washed once in 50% formamide, 2 x SSC, pH 7 at 45

C for 15 min, and finally in PN buffer at room temperature for 15 min. After draining excess liquid from the arrays, we mounted them in a solution containing 90% glycerol, 10% PBS and 1

M DAPI, and then sealed them with a cover slip. We drained excess glycerol-DAPI solution onto a paper tissue.
Imaging and analysis . We acquired 16 bit 1024x1024 pixel DAPI, Cy3 and Cy5 images using a custom built CCD camera system
1 , although we have also used commercial laser scanners for
7

this purpose . We used “UCSF SPOT” software (A.N.J. manuscript submitted) to automatically segment the spots based on the DAPI images, perform local background correction and to calculate various measurement parameters, including log
2 ratios of the total integrated Cy3 and
Cy5 intensities for each spot. We used a second custom program SPROC to associate clone identities and a mapping information file with each spot so that the data could be plotted relative to the position of the BACs on the September, 2000 freeze of the draft human genome sequence
(http://genome.ucsc.edu). SPROC also implements a filtering procedure to reject data based on a number of criteria, including low reference/DAPI signal intensity and low correlation of the Cy3 and Cy5 intensities with a spot. The SPROC output consists of averaged ratios of the triplicate spots for each clone, standard deviations of the triplicates and plotting position for each clone on the array, as well as other clone information stored in the database, such as STS content (Tables
A and B). We edited the data files to remove ratios on clones for which only one of the triplicates remained after SPROC analysis and/or the standard deviation of the log
2 ratios of the triplicates was > 0.2.
8

Web Figure Legends
Fig. A . Image of a HumArray 1.14 array after hybridization.
We hybridized Cy3 labeled HT29 cell line DNA (green) and Cy5 labeled normal male reference
DNA (red) to an array and imaged it using the CCD camera system. We produced the two color image by overlaying the individual Cy3 and Cy5 images using Adobe Photoshop.
Fig. B . Normalized copy number ratios for three comparisons of BT474 DNA and normal male reference DNA on an array of 1777 clones (HumArray 1.11).
In two hybridizations, we labeled the genomic DNAs by random priming (pink and blue) and in the third by nick translation (yellow). Data are plotted as the mean log
2 ratio of the triplicate spots for each clone normalized to the genome median log
2 ratio. The BACs are ordered by position in the genome beginning at 1p and ending at Xq. The borders between chromosomes are indicated with vertical bars.
Fig. C . Normalized copy number ratios for X chromosome clones for 5 female/male comparisons.
We found consistent ratio differences among the clones. Data are plotted as the mean log
2 ratio of the triplicate spots for each clone normalized to the genome median log
2 ratio. The BACs are ordered by position from Xpter to Xqter. Note that ratios on the Xp and Xq telomere clones are at the genome average, since these clones contain sequence shared with the Y chromosome.
9